Burkholderia pseudomallei, are commonly observed in Thailand and Australia (but also other equatorial regions), where mortality rates are 40% and 14%, respectively. The difference in these mortality rates has been attributed to healthcare quality (28), but our data demonstrate that pathogen populations also differ between these regions (145). In addition, prospective clinical studies in Australia are under way with promising preliminary associations between strain types and outcomes (Currie, Tuanyok, Wagner, Keim, et al., unpublished data). It is well established that B. pseudomallei contains an "open" genome (90) that recombines at a high frequency, leading to great diversity within and among pathogen populations. We believe differential virulence among pathogen populations (strains) contributes to differential mortality rates around the globe. Our primary hypothesis is that highly diverse B. pseudomallei strains have different virulence levels, and that these virulence differences will depend on the strain genomic composition (e.g., genomic islands). Multiple infection routes have been documented. Melioidosis infection routes are frequently hard to determine in the clinic, but inhalational and percutaneous routes both occur. Melioidosis incidence increases following tropical storms and near-drowning (30, 31), consistent with a pulmonary route. However, most melioidosis cases probably result from percutaneous inoculation (39), which is consistent with the presence of skin abscesses and dermal lesions (28). Virulence varies according to the infective route in animals and depends on the particular strain (see CK#3 and (11,146)). Animal models are important. Because human studies can be problematic, animal models are a common and powerful research approach to understand pathogen virulence. The mouse is the least expensive model, yet a very powerful one, but a single model may not always accurately represent diseases in other animals, including humans. Developing additional animal models (e.g., multiple mouse strains, rat, nonhuman primates) can support initial studies in the mouse and make our disease understanding more generalized and representative for human disease intervention. Knowledge and understanding of animal models is critical to infectious disease research.